feat: Slovak language support (sk-SK)

[braklo]

Summary

Requesting Slovak (sk-SK) language support in Harper. As far as I can tell there is no existing issue or PR tracking Slovak specifically, so this is to track it. See also the general non-English roadmap in #2654.

Why

Harper's speed, offline operation and privacy-first design make it very attractive for Slovak users — a niche currently served only by heavier, slower tools (LanguageTool / ltex-ls). There is no lightweight, offline grammar/spell checker for Slovak today.

Open resources to seed it

• Hunspell sk_SK dictionary (.aff + .dic), widely packaged on Linux distros.
• Slovak National Corpus data and the sk-spell project (by Zdenko Podobný), which is also the basis of LanguageTool's Slovak spellchecker. Permissively licensed (LGPL / MPL).

Notes on difficulty

Slovak is a West Slavic, strongly fusional language: rich nominal declension (6–7 cases × number × 3 genders, with animacy), extensive verb conjugation and aspect, and productive derivation. A plain word list won't be enough — solid morphological annotations / word-forms will matter. In terms of effort this is probably closer to the Belarusian attempt (#2195) than to the Romance/German PRs (#2150, #3402).

How I can help

I'm happy to contribute on the non-Rust side — assembling dictionary-sk + annotations from the resources above. Since #2150 and #3402 are still converging on the multi-language framework, I'd appreciate guidance on the currently recommended branch/structure to target before I start.

[hippietrail]

Basically we'd need a "champion" for the language support. Preferably somebody who can program but at least somebody willing and capable of editing and maintaining the dictionary.dict and annotations.json files. These are the basic components upon which all else is built.

[braklo]

@hippietrail Thanks — I'd like to volunteer as the champion for Slovak (sk-SK).

Full disclosure so there are no wrong expectations: I'm not a programmer. I'm a native Slovak speaker and I handle the technical parts with AI assistance. But I'm willing and able to build and maintain the dictionary.dict and annotations.json files long-term, and I have the language resources to do it properly — hunspell sk_SK, the Slovak National Corpus (Slovenský národný korpus), and the sk-spell wordlists.

Slovak is a heavily inflected Slavic language, so the dictionary side will be substantial — closer in scope to the Belarusian effort (#2195) than to the Romance/German work.

To start on the right foot: which structure/branch should I target? From the roadmap (#2654) the multi-language framework looks like it's converging between #2150 and #3402, and I'd rather build the SK files against whatever layout you'd actually merge than guess and redo it later. A pointer to the canonical dictionary.dict / annotations.json format (and any existing language as a template) would be enough to get me going.

[hippietrail]

To start on the right foot: which structure/branch should I target? From the roadmap (#2654) the multi-language framework looks like it's converging between #2150 and #3402, and I'd rather build the SK files against whatever layout you'd actually merge than guess and redo it later. A pointer to the canonical dictionary.dict / annotations.json format (and any existing language as a template) would be enough to get me going.

Those are both works in progress so you'd need coding skills to go that way. The most straightforward way would be to fork the current Harper and replace the English. Almost everything you'd work on would be the dictionary and annotations, and the guys working on the language support will be able to (relatively) easily integrate anything you give them. You'd probably want to look at what the others have done with their dictionary and annotations files to figure out how to get the dictionary to cover all inflected forms of all words and not accidentally any non-words.